Epigenetic modulation of immunotherapy and implications in head and neck cancer

Immunotherapy has emerged as one of the most exciting clinical frontiers in cancer management over the latest decade and has given new hope to poor prognosis cancer patients. Immune checkpoint blockade, including inhibition of the programmed death 1 (PD-1) receptor pathway or CTLA4/CD152, aims to reinvigorate the host anti-tumor immune response. Blocking these pathways results in an effector T cell response leading to cancer cell eradication [1‐3]. Despite impressive results from multiple clinical trials demonstrating a durable response, only a fraction of patients in most cancer types are responsive to immune checkpoint blockade treatment [4‐9].

Head and neck squamous cell carcinomas (HNSCC) represent one of the tumor types where up to 80% of patients do not respond to anti-PD-1 based therapies. HNSCCs are the sixth most common cancer worldwide and can be divided into a classical carcinogen (tobacco and alcohol) induced variety and one where the human papillomavirus (HPV) is the primary etiology. Both HPV-positive and HPV-negative HNSCC tumors are highly immune infiltrated [10]. The HNSCC mutation burden is at the relatively higher end within the spectrum across human cancer types, which usually predicts a higher number of mutation derived neoantigens, the ultimate target of the immune system [11, 12]. The use of PD-1 blockade with nivolumab or pembrolizumab in second-line recurrent and metastatic HNSCC was approved by the US Food and Drug Administration (FDA) in 2016 [13]. Recent data from KEYNOTE-048 have expanded the use of pembrolizumab to first-line recurrent and/or metastatic HNSCC. However, the response rates of only 15–20% highlight the need for continued investigation to advance therapeutic options [14‐16].

Circumventing resistance mechanisms to immunotherapy represent a robust research area that is the critical barrier to improve the clinical outcomes of cancer patients [17, 18]. Mechanisms of resistance can be divided into (a) tumor cell intrinsic immunoevasive genetic pathways, (b) a suppressive tumor microenvironment, or (c) host factors including somatic variants and the gut microbiome [19]. A major cancer cell resistance mechanism involves lesions in genes in the interferon-gamma (IFNγ) signaling pathway or antigen presentation machinery that independently impair the efficacy of immunotherapy [20‐23]. High infiltration of immunosuppressive regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs) in the tumor microenvironment is associated with a poor prognosis in various cancers [24, 25]. Extending the work in mouse models [26], a correlation between the gut microbiome and efficacy of immune checkpoint therapy has been shown in melanoma, lung, and kidney cancer patients [27‐29].

Various combinatorial therapies have been proposed to improve the response rate and overcome resistance to anti-PD-1 treatment. These strategies include combining with other immunotherapies, chemotherapy, targeted therapeutics, or radiation [30‐40]. Several recent negative trials have unfortunately given pause on next steps in combination approaches. The EAGLE Phase III trial that evaluated combining durvalumab with tremelimumab versus standard of care chemotherapy failed to show an enhanced clinical impact in recurrent/metastatic HNSCC [41]. The Javelin Head and Neck 100 Phase III trial combining avelumab with chemoradiotherapy in the definitive HNSCC setting was closed early as there was no enhanced activity relative to chemoradiation [42]. These negative results of Phase III trials have highlighted that exploration of rational combinations beyond existing therapies is needed.

The ideal therapy to combine with anti-PD-1 agents would be one that has pleiotropic effects on multiple targets that together would limit development of resistance. As such, therapeutics that target epigenetic modifications of both cancer cells and components in the tumor immune microenvironment represent such an opportunity. In this review, we focus on recent preclinical studies with epigenetic modulation in cancer immunotherapy and discuss the effects of epigenetic interventions in cancer cells, T-cells, natural killer cells, and macrophages. We review the potential of existing epigenetic therapeutics in promoting antitumor immunity followed by a brief discussion of strategies to define new therapeutic targets. Finally, we describe the rationale for combining epigenetic targeting and immunotherapy to improve the clinical outcomes of HNSCC.

In broad terms, epigenetic alterations refer to gene expression changes resulting from effects of DNA methylation, histone modification, regulatory non-coding RNA, or transcription factors and not caused by alteration in DNA sequences. Cancer cells are highly enriched for epigenetic abnormalities, which makes epigenetic regulators attractive targets in cancer treatment. DNA methylation, histone acetylation, and histone methylation, specifically EZH2-mediated histone H3 lysine 27 tri-methylation (H3K27me3), represent the major targets where therapeutics is available for epigenetic therapy.

The inhibition of DNA methyltransferases (DNMTs) targets promoter region hypermethylation on the cytosine residues. DNMT inhibitors (DNMTis), such as 5-azacytidine and 5-aza-2′-deoxycytidine, have been approved by the United States Food and Drug Administration (FDA) in the treatment of patients with myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) [43‐45].

Post-translational histone modifications include histone acetylation and methylation. Histone acetylation increases the accessibility of chromatin to the transcriptional machinery by removing the positive charge on the histones. Blocking of HDAC activity using small molecule inhibitors has shown potent effects in suppressing tumor progression and apoptosis of cancer cells. HDAC inhibitors, such as Vorinostat, Belinostat, and Panobinostat have been approved for the treatment of cutaneous T cell lymphoma, chronic lymphocytic leukemia, and multiple myeloma, respectively [45, 46]. Vorinostat, also known as suberanilohydroxamic acid (SAHA), is one of the most advanced small molecule pan-HDAC inhibitors, which is administered orally [47].

Enhancer of Zeste 2 Polycomb Repressive Complex 2 Subunit (EZH2) is the catalytic component in the Polycomb Repressive Complex 2 (PRC2), which tri-methylates lysine 27 of histone H3 (H3K27me3). The inhibition of EZH2 activity may slow tumor growth by upregulating tumor suppressor gene expression. The clinical efficacy of an orally administered EZH2 inhibitor, tazemetostat, is under investigation in multiple clinical trials including non-Hodgkin lymphoma (NHL), INI1/SMARCB1-negative tumors, synovial sarcoma, colorectal cancer, prostate cancer, renal cell carcinoma, ovarian cancer, and mesothelioma. Tazemetostat has recently been approved for the treatment of metastatic or locally advanced epithelioid sarcoma and relapsed or refractory (R/R) follicular lymphoma (FL) whose tumors are positive for an EZH2 mutation.

The goal of cancer immunotherapy is to modify the interface between cancer cells and immune compartments in the microenvironment to support an anti-tumor response. We first discuss recent preclinical studies on epigenetic modulation in cancer cells that may facilitate the antitumor response by impacting innate and adaptive immune pathways.

A significant subset of immunotherapies, including immune checkpoint blockade, seek to directly modulate T cell function to eradicate cancer cells. An effective anti-tumor T cell response requires TCR signaling pathway stimulation, clonal expansion, and differentiation to effector functionality, in which epigenetic modulation plays a key role. Therefore, the idea of therapeutically targeting T cell epigenetic programs is of great interest in achieving better clinical outcomes.

Tumor associated macrophages (TAMs) are distributed in a polarization spectrum from classically activated M1- macrophages (M1s) to alternatively activated M2-like macrophages (M2s). M1s and M2s have opposite functions in tumor progression: M1s are pro-inflammatory tumor-inhibiting, while M2s are immunosuppressive tumor-supporting [79]. High infiltration of M2s correlates with poor prognosis of HNSCC [80, 81]. Therefore, repolarization of TAMs towards M1s represents a major strategy in manipulating TAM to control tumor progression. Interferon gamma (IFNγ), lipopolysaccharide (LPS), and other Toll-like receptor ligands favor M1 polarization. Ivashkiv and colleagues demonstrated that IFNγ induced EZH2-mediated histone H3K27me3 modifications at the promoter regions of anti-inflammatory genes suppressed the gene expression in primary human monocytes-derived macrophages. In addition, IFN γ-induced H3K27me3 resulted in the stable suppression of gene expression [82]. Therefore, EZH2 is involved in the chromatin remodeling process induced by IFNγ to repress anti-inflammatory gene expression in macrophages to achieve and maintain the activated M1 state [82].

SOCS1 (suppressor of cytokine signaling 1), a key cytokine signal negative regulator has been shown to be under epigenetic modulation in macrophages [83]. Cheng and colleagues found that DNMT1 mediated DNA methylation in the promoter region of SOCS1 resulted in the suppression of its gene expression. Inhibition of DNMT activity using 5-aza-2′-deoxycytidine in LPS-activated RAW264 macrophage cells reduced SOCS1 expression and consequently enhanced the production and release of pro-inflammatory cytokines such as TNFα and IL-6 [83].

Further granular details about the interaction between tumor and host immune compartments are being revealed in expanded studies of cancer immunotherapy response including by the use of innovative technology platforms. Together, these approaches will provide invaluable information in developing new therapeutic targeting strategies.

For example, the anti-tumor contribution of tumor infiltrating B cells has been controversial [84‐87]. Recently, the association between the presence of B cells in tertiary lymphoid structures in tumors and favorable responses to immunotherapy has been demonstrated in soft tissue sarcoma, metastatic melanoma, and renal cell carcinoma, respectively [88‐90]. Thus, different B cell subsets in the tumor microenvironment may have distinct contributions in an antitumor immune response [91]. The effect of epigenetic modulation in B cells to cancer immunotherapy response remains to be explored. However, the differentiation and activation of B cells are under the regulation of histone and DNA modification, such as H3K27me3, H3K9me3, and DNA methylation [92]. It will be of interest to investigate the effect of epigenetic manipulation on the spatial distribution and functional contributions of B cell subsets in tumors.

Development of CRISPR/Cas9 systems for ablation of specific gene function in large-scale discovery screening has allowed unbiased interrogation of critical immune pathways. Using genome-wide CRISPR/Cas9 functional screens, several studies reported tumor cell intrinsic factors in immunotherapy resistance pathways in models of melanoma and leukemia [58, 93‐95]. From these studies, chromatin structure modulators such as Ezh2 and PBAF complex components were revealed to be integral to the tumor cell/T cell interaction. Studies in multiple cancer types have validated the effect of Ezh2 on MHC class I antigen presentation [57, 58, 96], including HNSCC [56]. In contrast, the role of PBAF (Polybromo-associated BAF complex) in tumor cell interferon gamma sensitivity was only demonstrated in melanoma models [93]. Although further studies are needed, these findings suggest both cancer-type specific and conserved common epigenetic-related immune modulators exist, and defining their respective relative contribution will impact therapeutic target identification.

To specifically identify epigenetic modulators involved in the sensitivity or resistance of cancer immunotherapy, an epigenetic-focused in vivo screen was also performed using lung adenocarcinoma models [97]. Compared with in vitro screens, in vivo screens provide relatively more physiological conditions with endogenous antitumor immunity and immunotherapy treatment as selection pressure. In addition, in the complex tumor microenvironment, the interaction between cancer cells and immune cells is also not limited to “two-cell-type” coculture of the in vitro screening system. On the other hand, in vivo screens require large numbers of mice especially with unbiased screening libraries, which can be labor intensive and cost prohibitive compared to in vitro screens [95]. In vivo T cell CRISPR screens have also been shown to be feasible in immunotherapy target discovery [98]. Therefore, CRISPR/Cas9 screens using HNSCC models represent a valid strategy for identifying immune modulators involved in the resistance or sensitivity pathways of HNSCC to immunotherapy.

As discussed earlier, despite the durable response demonstrated with immune checkpoint inhibitors in patients with recurrent/metastatic HNSCC, various combinatorial treatment strategies, including with epigenetic targeting, are being actively tested to facilitate the antitumor immune response. Rodriguez and colleagues completed a Phase II trial of pembrolizumab and vorinostat in two distinct head and neck tumors-recurrent/ metastatic HNSCCs and salivary gland cancers [99]. For the HNSCC cohort, grade 3 AEs of any cause were observed in nine (36%) patients, which is higher than that reported for pembrolizumab alone. Although there were several limitations, including cohort size and heterogeneity and that there was no run-in phase of vorinostat alone, there was a 32% overall response rate that warrants further exploration. A second ongoing clinical trial is assessing whether addition of azacitidine to a durvalumab/tremelimumab combination will be safe and improve outcomes in recurrent/metastatic HNSCCs who have progressed on anti-PD-1, anti-PD-L1, or anti-CTLA-4 monotherapy (NCT03019003). There are several trials integrating EZH2 targeting with checkpoint inhibition (for example NCT03525795, NCT03854474) but these have not expanded to include HNSCCs. Our preclinical work represents a rational basis to complete such a study. Future studies may also integrate targeting of specific protein methyltransferases/demethylases that are associated with the immune-cold phenotype of HPV-negative HNSCC and could be considered for preclinical investigation to decipher mechanisms of CD8+ T cell exclusion in this disease (PMID: 29348866). Finally, patient selection with epigenetic biomarkers to define susceptible patients should be a goal of future clinical trials.